Ask HN: Is AI generated artwork my artwork? | Hacker News Reader

Ask HN: Is AI generated artwork my artwork?

3 pointskixpanganiban14 comments

I've been generating images with AI for a while. First with Nightcafe Studio, then Disco Diffusion, and eventually settling down with Midjourney. I love the images I generate from MJ, especially after I've figured out the techniques.

For context, I can spend hours generating a single image that I like, continuously refining my prompts and word weights, stylization, composition, and so on. It certainly takes skill, but obviously very different from drawing the entire thing myself (which I can't).

I liked one of my recent images so much that I decided to post it publicly on an online forum for the first time. The response I got was very polarized. Among the standouts were: "stop claiming AI artwork to be your artwork, unless you made the AI that generated it." I disagree, obviously, but it got me thinking. What do you folks think?

I think that the democratization of art through AI is revolutionary, but I don't think we've figured out how we treat the ownership of artwork from AI yet. It learns to make images much the same way artists learn from other artists, but does that count?

Yes. It's your art. AI tools are just tools. When you use Photoshop, no one says Photoshop made the images. The art you make with AI software did not exist until you wrote the prompts. It isn't a copy of anything.

This stuff is all new and will take time to settle in but I have no doubt as a user of Photoshop since v1.0 and DALL-E since the beta -- that these AI tools will be the norm in every artist and designer's toolbox.

My thoughts exactly. I remember the early days when the photography world was just as split with Photoshop, where if a person used Photoshop to make edits to their photograph it is no longer "authentic" or "real". Nowadays, just about any Bob and Betty and their uncle uses filters on any social media platform and no one contests that the photo is theirs.

I can see a world where DALL-E and SD, or at least versions of them, are part of Photoshop (or some other photo editing suite). Much like how the clone tool and content-aware fill were groundbreaking, a new tool where you can type in "replace background with a dystopian city, cyberpunk style" would be just as ubiquitous (and reminds me of this post by Runway ML: https://twitter.com/runwayml/status/1568220303808991232?s=20...)

I've recently gotten a lot of attention for my Midjourney stuff, and I did cause it to be, and I do own the copyright, but I don't claim it as "my art." I think we know that's disingenuous because we didn't create the specific composition.

It's like if I asked one of my assistants to paint something based on a paragraph I wrote, is that MY art? Not at all.

The "we didn't create the specific composition" bit is the one that's most contentious, I think. Even in Midjourney, getting an image to be just right takes a lot of time and skill with all the flags, word weights, and specific word prompts you have experiment with over hours. To me, the resulting image feels very much like my composition, with every element of the image placed exactly where I want them on the canvas, with the specific style, color, and feel that I dictate.

Your analogy with an assistant is interesting, but to me is flawed. A better analogy in my view would be an art director or concept artist coming up with a concept that they ask their team to execute. In that case, it's normal for the art director to take credit for the art, and in this context where the AI is just a tool with no agency, I feel the same way.

You have a point. Can you share a prompt with that level of specificity? Its fascinating to see this new sort of science of prompt engineering.

An example prompt for a recent image I generated: "skull::0.75 of a woman with long hair, ornate::0.9, hooded::0.7 robe::0.7, highly detailed::0.8, decorated, full body, standing in a mystical::0.8 forest::0.7, style of Aleksi Briclot, realistic::-1 photoreal::-0.7, --no water --no sky, --ar 4:5"

Prompt aside, the final image took me about an hour to make, rerolling a number of times to get the composition I want with the base v3 algorithm, then upscaling and then remastering with the --test --creative --upbeta mode, and rerolling again until I got the final remaster just right.

It looks great, and there's definitely an art to prompting. But whether you created the exact composition as we were discussing still isn't solved for me. It's not full body, there's barely something that looks like a skull, and it's off to the side.

It's a very interesting question that I think we'll be debating and thinking about for some time.

If you never tell anyone it was AI that generated it, it's yours.

I don't think Frank Herbert can take credit for the imagery in Villeneuve's Dune, even though he came up with the text prompts. He gets full credit for the text prompts, though!

No it's not. It's debatable if it even qualifies as 'art'.

Who owns the copyright?

Why do you think so?

Would the artwork be in its own category: AI Artwork? Otherwise, I find it hard to support an AI-Artist who machine generated artwork based on "stealing" and "feeding" it 100.000s of creations from traditional/digital artists. The potential to profit off of it in one form or another is un-ethical. AFAIK, nobody have the rights to use their artwork without permission. Artists' artwork is not under a MIT license nor is it for sale as a stock image.

It's flawed to think an artist referencing others to produce their original is the same as an AI-Artist stealing artwork to generate an original. Youtube, Facebook, TikTok reciprocates value to its customers in exchange for selling their data to advertisers.

What value are YOU, the AI-Artist, providing the artist for their data?

Note: I'm ignorant in this area. My views are based on personal conversations with how influencer artists' feel about this popular trend.

I'm genuinely curious why you think the AI "steals" art from artists. That train of thought seems to imply that any material generated by any machine learning process is "stolen" from the training data. Why do you think that's the case?

Speaking of data and profits, if I were a digital artist with the ability to hand-draw images, and I decided to draw a unique composition that follows say the art style of Anne Stokes, do you think it would be the case that I "stole" from Anne Stokes?

It's unclear to me why you point to the need of me, the AI-Artist, to deliver some obscure "value" back to the artists from whom training data was referenced. The only value I have the ability to deliver is to the AI model and its authors, other than perhaps propagating the art style or work of the authors of the art basis.

Reply on news.ycombinator.com